P4P flies forward into tree during active track
1219 7 2017-5-30
Uploading and Loding Picture ...(0/1)
o(^-^)o
Elevate72
lvl.1
Flight distance : 231572 ft
Switzerland
Offline

Hello everyone.
Some days ago i used my brandnew P4P to film a boat by following it with the active track mode. Light and variable winds, no control inputs from my side, the drone followed the boat along the river. Suddenly it increased the speed to 55kmh ( I thought the limit for P mode was 50kmh... ) and started a light drift to the right. As I was on the boat the distance of the drone to the trees on the shore was hard to judge. No warnings were triggered before the crash and it flew straight into the trees at 16m height. It flipped and luckily recovered at 12m.
After that the obstacle avoidance was active and triggered several warnings.

Why the foreward sensors did not trigger the obstacle avoidance? I did some test flights afterwards and everything seems to work allright.

Now I am not sure if I should send it back to DJI to have it checked or if this was normal behaviour in this situation....
Please check out the movie and the screenshots of the flight data just before and after the crash.

Thanks for suggestions!!

https://www.dropbox.com/s/et9ze0inej2l22k/P4P%20flies%20into%20tree.mp4?dl=0




Foto 29.05.17, 14 24 26.jpg
Foto 29.05.17, 14 25 59.jpg
Foto 29.05.17, 14 28 51.jpg
2017-5-30
Use props
eColumbia01
lvl.4
United States
Offline

From looking at the video, the drone hit a branch as opposed to a solid wall. I'm not sure how good the obstacle avoidance is and i have not tested this. For instance, if the drone is flying directly towards a tree with lots of branches and leaves, then the sensors work and stops the drone. However if the drone is flying at an angle to the tree, then I'm not sure if the obstacle avoidance is accurate enough to stop the drone. I also have not tested this.

On another note - you are very lucky the drone did not drop into the water. Personally I would test out the drone if all seems well, then continue flying. It is probably a good idea to reset the IMU and gimbal.
2017-5-30
Use props
Mobilcams
lvl.4
Flight distance : 2907674 ft
United States
Offline

One thing I noticed is that it looks like active attack lost the boat as it's subject. I have noticed strange behavior from my P4 and my P4P using this feature and use it only occasionally. Just a chance you take when letting it fly autonomously.
2017-5-30
Use props
Aardvark
First Officer
Flight distance : 384432 ft
  • >>>
United Kingdom
Offline

There is a specific switch to toggle obstacle avoidance on/off for active tracking in DJI Go 4 in the 'visual navigation Settings' menu. Could this have been off ?

Edit:- had it been off then the aircraft should still come to a stop and hover...
2017-5-30
Use props
Elevate72
lvl.1
Flight distance : 231572 ft
Switzerland
Offline

Thanks for your inputs.

I did a IMU, Compas and Gimbal calibration and it seems to work perfectly now. As it did before the crash...
The drift angle is very small, so i think the forward sensors with its 60 degree angle should have worked.

I agree it must have lost the subject, but the warning about it only comes in after the crash ( see warning log. crash is at 16:43:18...)
Only after the crash obstacle avoidance is active.
Why did it drift right? And why did it speed up to 55kmh?

I have all the avoidance features in the visual navigation settings turned on...

2017-5-30
Use props
Geebax
Captain
Australia
Offline

Looking at the footage two things come to mind. First, obstacle avoidance is a bit iffy when dealing with tree foliage, and secondly, the speed was probably too fast for it to react in time. I must say though, you are a braver man than I, I would not be using an autonomous mode like that over water.
2017-5-30
Use props
DJI-Jamie
DJI team
Flight distance : 112405 ft
United States
Offline

Branches can be tricky, as the thin ones are difficult to distinguish for the sake of depth perception. The only thing I could see, however, is that the OA could have gotten confused due to the sudden change in lighting. Regarding the initial veer off, it could have been that the active track function lost its subject due to the bright white boat blending in with the reflective water in the distance. It was very fortunate that the aircraft didn't fall into the water. If the issue persists (preferably on land), I would recommend sending it in for evaluation.
2017-5-30
Use props
Elevate72
lvl.1
Flight distance : 231572 ft
Switzerland
Offline

Thanks Jamie.
That sounds like a plausible explanation. In this case it doesn't make sense to send it in for evaluation, as it is working perfectly now.
2017-5-31
Use props
Advanced
You need to log in before you can reply Login | Register now

Credit Rules